Patients with early AD exhibit impairments in the monitoring and control of cognitive processing (i.e., executive dysfunction [ED]). Recent evidence suggests that the appearance of ED in elderly persons with mild cognitive impairment (MCI) may be a precursor of dementia. However, executive function encompasses several different, dissociable, cognitive processes, and the relevance of particular executive deficits to functional decline has not yet been established. In addition, MCI has been variably defined, and precisely which patients are at high risk for functional decline leading to dementia is not yet clear. Therefore, the primary goals of this study are: 1) to determine the specific executive domains that are impaired in patients with amnestic and nonamnestic forms of MCI; 2) to determine the extent to which particular executive impairments affect everyday functioning differentially in these groups; 3) to determine whether ED among subtypes of MCI is predictive of subsequent functional decline, measured by the sum of boxes from the Clinical Dementia Rating scale (CDR-SB); and 4) to test a factor analytic model of executive functioning in normal elderly and those with MCI. Patients with amnestic MCI, nonamnestic MCI and normal control participants will be recruited. They will be studied with 18 cognitive tasks that are proposed to assess six conceptually distinct domains of executive function: 1) inhibition of prepotent responses; 2) decision-making and judgment; 3) planning and sequencing; 4) concept/rule learning and set shifting; 5) spontaneous flexibility and generativity; and 6) working memory and resource-sharing. Confirmatory factor analysis will be used to test and validate this cognitive model, which will be modified if necessary. Questionnaires assessing subjects' everyday functioning will be completed by knowledgeable informants to assess the impact of ED in daily life. The executive factor scores of the two MCI groups and the normal control group will be compared with regression models for multivariate outcomes. They will also be examined as predictors in regression models of scores on the Dysexecutive Questionnaire and the ADCS/MCI-ADL scale. All participants will have annual clinical re-evaluations, and incident cases of dementia will be recorded. Two years later, all available subjects will be re-examined with the executive function tasks and questionnaires. The re-test data will allow us to determine the progression of particular executive impairments and their predictive validity as pre-dementia markers of functional decline (CDR-SB score) in amnestic and nonamnestic MCI.